Apr 04 2014

How Would Exascale Computing Impact K–12?

This next-generation, high-performance computing technology looms large as a game changer for education.

Calculators revolutionized math education by allowing students to scrap the grunt work of multiplying and dividing so they could focus on higher-order thinking.

Then along came the Internet, giving them access to nearly limitless information.

Eventually, students may even be able to conduct advanced research studies with just a few keystrokes.

“The information that people will be able to access will be phenomenal,” predicts Earl Joseph, program vice ­president for high-performance ­computing (HPC) at IDC. “You’ll have fourth-graders who are able to do surveys of almost everybody on Earth — research that, at this point in time, would amount to a dissertation. I think it’s going to redefine how our education system is going to work.”

Moving Beyond the Petascale Barrier

The technology that has Joseph so ­excited is exascale computing, a system that’s capable of a quintillion floating-point operations per second (or “flops”) — a thousandfold increase over the petascale barrier that was ­broken in 2008.

At that speed, he explains, computers will be able to generate complex answers in response to simple searches, giving students the equivalent of a team of Nobel laureates to help them answer questions. Joseph says exascale supercomputers also will enable the sort of data processing necessary to render rich virtual reality environments for a whole classroom of users simultaneously. The numbers will be crunched at the supercomputer, he says, and then the images will be sent to devices worn by students (meaning schools can benefit without owning their own supermachines).

The theoretical peak performance, in petaflops, of Tianhe-2, making it the world’s most powerful supercomputer

SOURCE: TOP500 (November 2013)

“Once you get to that level, then you can go to impossible places — under the ocean, to Mars, inside a volcano,” Joseph says. “Someone in fourth or fifth grade will be learning science in a way that even scientists can’t do today.”

Outside of education, exascale ­computing is expected to spur advances in many industries and research areas, including climate modeling, aircraft design, medicine and financial security.

“The question of why you build a more powerful supercomputer is similar to asking why you build a more powerful microscope,” says Addison Snell, CEO of Intersect360 Research, which specializes in HPC. “You’re trying to make better tools for scientific discovery.”

Most experts predict that exascale computing will become a reality sometime between 2018 and 2022, says Jack Dongarra, a distinguished professor of computer science at the University of Tennessee and an expert in supercomputing.

“It’s going to be a challenge to produce a machine that’s power-efficient enough to run exascale,” Dongarra continues. Using existing technology, the electric bill for an exascale computer would be $200 million a year.

Whenever scientists do build an exascale computer, one thing is certain: They won’t stop there.

“People get really excited about the next prefix, like ‘peta’ or ‘exa,’ ” Snell says. “These are road signs on a highway. They’re not goals in themselves. Until we reach the end of science, we’re never going to stop pursuing that next level of scalability.”